The long-term objectives ofthis NRSA are 1) to develop substantive expertise in the study of child development to complement existing methodological skills and ongoing methods training, and 2) to apply advanced analytic techniques to study the effects of maltreatment on adolescent social and psychological development. The training sequence for meeting these objectives includes: 1) participation in a weekly developmental seminar and consortium with scholars in the field of child development, 2) academic and professional supervision by an advisory panel of experts in the fields of child development, child maltreatment, and quantitative methods, 3) participation in small multi-institution and multi-disciplinary maltreatment research groups, 4) mentorship in child development and methodological techniques particular to the field of child development, and 5) coursework in child health and development. In conjunction with and to facilitate these training objectives, the specific research project undertaken during this training is made up of three focused specific aims. Aim 1 is to develop measurement models of maltreatment (abuse, neglect) using continuous and discrete latent variable models with a variety of maltreatment indicators in a population based sample of adolescents engaged with US child protective services (CPS). Aim 2 is to assess the effects of child maltreatment as measured in Aim 1 on the trajectories of youth well-being in the domains of anxiety;depression;withdrawal;delinquency and aggression;and, social, thought, and attention disorder. Models used in this assessment will be advanced within a developmental-ecological theoretical framework. Aim 3 is to investigate whether participation in formal and informal social organizations outside of the family moderates the effects of maltreatment on child development. Participation in extra-familial social organizations is hypothesized to be protective. A primary contribution ofthis research will be to disseminate higher quality measures of child maltreatment from survey data using structural equation modeling with latent variables. Random measurement error and in some cases, response bias, will be accounted for in measurement models. In addition, models of relationships will be holistic in that they will involve multiple contexts and change over time, the most recent methods for the analysis of complex samples will be used to obtain consistent population estimates. This research addresses the social and affective development;child maltreatment and violence program area ofthe NICHD. The advanced modeling and inference to populations using complex samples is applicable to all public health research that seeks unbiased measurement of population parameters.